An engineer to choose your support AI and make it work in your helpdesk. Sierra, Decagon, Intercom Fin, and your helpdesk's own AI all publish strong case studies. We put a senior engineer on your team who tests them on your tickets, picks with you, and does the integration, escalation, and QA work that decides whether the numbers hold.
Bring a 90-day ticket export from your helpdesk.
Published results in customer support
Three published results. The first shows the careful version, and the last shows what happens when volume is the only goal.
WeightWatchers' Sierra agent contained nearly 70% of cases in its first week, with CSAT above 4.5 out of 5.
Started on a small slice of app traffic. Connected to internal systems so it could take actions. IT and security attacked it with adversarial prompts before launch.
Intercom's Fin resolved 55% of Synthesia's support conversations while monthly requests grew from 40,000 to 316,000.
The team estimated it would otherwise have needed to grow to 150 people. It handled the growth without growing the support team.
Klarna's OpenAI-based assistant handled 2.3 million conversations in its first month, two-thirds of its customer service chats.
In May 2025 Klarna's CEO said quality had suffered and that the company would hire more human agents. Volume alone was not the finish line.
Figures as reported by the vendor or customer at the linked source. They are not OpenNash results.
Where our engineer comes in
WeightWatchers started on a slice of traffic and had IT and security attack the agent before launch. Klarna later said quality had suffered. The difference is the work around the product. Our engineer takes that on, with a helpdesk admin seat, billing access, and a weekly review with your QA lead.
| Decision | What we look at | Outcome |
|---|---|---|
| Sierra, Decagon, Fin, or a custom build | Per-resolution price at your volume over three years, how many tickets need actions in billing or orders, and who owns the knowledge base and test set afterwards. | A written comparison with the cost at your volume. |
| Which ticket class goes first | 90 days of tickets grouped by intent, with volume, handle time, and reopen rate. | A ranked list. You choose. |
| What the agent may do alone | Refund and credit limits, account changes, and legal or chargeback language. | Rules written by your team, enforced in code, and tested on every release. |
| When it goes live | Shadow-mode drafts compared with what your agents actually sent. | Only after the pilot tests pass. |
Why forward-deployed engineers
Built for their largest accounts, deploying their own models. OpenAI's deployment company launched with more than $4B. ↗
Configure their own product, priced per seat or per resolution.
A small senior team, ex-Palantir, Meta, and Databricks. Any model, any vendor. Free pilot, then a flat monthly fee.
Even OpenAI closes the gap with engineers in the room. Source: OpenAI, What FDE means at OpenAI, Oct 2026.
The first eight weeks
Written down with your team in week 1. The pilot is free, and it has to pass these before anything goes further.
Example acceptance tests, not results from a customer deployment. Yours are written with your team.
Systems we work in
Rules from day one
FAQ
Possibly. We test them on the same set of your past tickets and price them at your volume over three years. If one wins, we do the integration and QA work. If none fits, we build, and you own the code and test suite.
No. Tickets, macros, and reporting stay where they are.
Your team. We write the escalation rules with your support ops lead, turn each one into a test, and run the tests on every release.
Why you need one
A senior engineer joins your team on a laptop you issue, tests the options on your data, and puts one customer support workflow into production in 4 to 8 weeks. It starts with a free 14-day pilot judged against test cases you agree in writing.
After launch, we run and improve the workflow on a flat monthly fee, or your team takes it over. The code, tests, and runbook are yours either way.
Bring a 90-day ticket export from your helpdesk.